Hewlett-Packard sort of announced its data warehouse appliance (if that is the right term—see later) last autumn. The company refers to this as a soft launch in the sense that they didn't talk much about it except to a few beta clients. Well, now it has had its hard launch (which is officially version 2.0) and the company is …
A bit more work to do at HP
While the specifications of NeoView are qute interesting, there are still a few things HP has to get the ticks in the box on. First, demonstrating some massive workloads - in the 100TB+ with thousands of updates per second. With Mark and Randy at the helm, this is an obvious chore to get completed: how NeoView will fare in a head-to-head with a big Teradata installation - especially one that has been going for some time - has yet to be seen.
Second, handling complex DSS reporting, both from tools and "hand-crafted" reports. These are the stock-in-trade of the typical large Teradata installation, especially one that has been running for more than a few years. Teradata has some killer abilities with fully-normalized and near-normal data in huge complex joins - and TDAT customers DO take advantage of this type of query in daily operations. Again, nothing the the leaders don't know about, but still something that has to be passed.
Third, will HP provide a reasonable migration path for some of the key Teradata and DB2 accounts? HP is actively recruiting former Teradata folks for the new services support for NeoView, so familiarity with the competition and the complexities therein will be available. However, "the best intentions" and all that: migrating a large Teradata or DB2 installation is not something for the faint-hearted. If there is a need to keep TWO massive data warehouses running at a customer site (both the encumbent Teradata and the new NeoView) for anything over a year, there could be some issues.
Finally, there is the market for massive databases itself. Teradata and IBM have survived the purge of nearly all the competitors back in the 1990s, and, even with appliances making some inroads into the "big iron" DW market, it's still a rarified customer base that HP is looking to enter. Even Teradata has been "down-scaling" its sales into smaller customers (the 1-4 node installations with a few tens of terabytes) because there is only so much revenue at the top. Of course, at this lower end you also have Microsoft, Oracle and the appliance vendors. Quite a crowd here, and, even more interesting, HP may have to canabalize some of its own sites, moving Oracle or SQL Server on HP hardware to NeoView on HP hardware. It's still a "win", but one of those "wink-wink nod-nod" deals that the hardware vendor can make by juggling margins and pricing within their product lines.
So HP will have a bit of work cut out for them. I personally do not expect to see "big things" out of the HP camp for at LEAST two years: the purchase cycle of the realy key customers will take that long to complete. I DO expect HP to start moving some targeted smaller systems into the Fortune 100 realtively soon; "beach head" systems, pretty much "freebies" along with some heavy-duty free services support to "shake the tree" and get both visibility AND some inside information on just what needs to be moved out of the Teradata or DB2 warehouse in order to take over the account.